DESIGNING NONLINEAR CONTROL SYSTEMS BY STATE-SPACE FLOW GRAPH OPTIMIZATION

This paper presents a method for planning optimal control for nonlinear processes by using discrete optimization algorithms. The principal concept is to analyze the autonomous dynamics and to use its properties in the design of an optimal state-space trajectory. The analysis is performed in a multi-stage procedure, based on a proper decomposition of an operational subspace into a set of segments. The arrangement of these segments is transformed into a flow graph structure. Flow values characterize the cost of driving the operational point between two adjacent nodes. A discrete optimization algorithm is used to search the state-space graph for an optimal path.